In inverse problems, we try to recover data from indirect measurements resulting from an ill-posed forward process. In the last few years, deep learning has shown great abilities to enhance such reconstruction techniques. Topics of interest include, learned regularisation schemes, statistical and optimisation approaches such as generative priors, plug-and-play methods, unrolled networks and neural operators for solving PDEs. Additionally, we are interested in uncertainty quantification for such reconstruction methods, for instance via Bayesian methods. Applications include problems from medical and non-medical image and signal processing, physics and engineering.
In this meeting, from 7-9 July 2025, in Bath, we aim to bring together experts from the fields of inverse problems and deep learning focusing on theoretical and practical questions.
Apart from the keynote talks, we have an open call for contributed talks and posters.
We are now inviting submissions for this conference.
You are invited to submit a proposal for a talk or poster that fits within the theme of the conference. Topics of interest are anything on the intersection of inverse problems and deep learning, such as learned regularisation schemes; generative priors; plug-and-play methods; unrolled networks; and neural operators for solving PDEs. Additionally, we are interested in uncertainty quantification for such reconstruction methods, for instance via Bayesian methods. Applications include problems from medical and non-medical image and signal processing, physics and engineering.
Please complete the form linked below, and indicate whether you have a preference for a poster or talk. Please note that we may not be able to accept all submissions. We may suggest that you bring a poster, rather than give a talk, so please indicate on the form whether you would be open to this.
We can receive submissions until Friday 28 February 2025. In mid-March, we will let you know the outcome of your submission and invite you to register for the conference. Registration will subsequently be opened for general admission. There will be a small fee of £50 payable for registration, to cover costs.
Register your submission using this form:
Please get in touch if you have any questions, by emailing maths4dl@bath.ac.uk.
General registration will open in April. There will be a small fee of £50 payable for registration to cover costs.
In the meantime, if you have any questions about the conference, please email maths4dl@bath.ac.uk.
The Maths4DL conference is being organised by:
Alexander Denker (UCL), Matthias Ehrhardt (University of Bath), Johannes Hertrich (Université Paris-Dauphine) and Hok Shing Wong (University of Bath).
